In the manufacture of internal combustion engine cylinder blocks, automotive heat exchangers, such as radiators or heater cores, or automobile fuel tanks and the like, it has long been necessary to test the same for leaks before approving them for use. Various approaches have been taken to this leak testing, some of which are shown in the following U.S. Pat. Nos.:
2,874,566 PA1 3,221,539 PA1 3,457,775 PA1 3,893,332 PA1 4,047,423
Despite these and other prior efforts, the need remains for leak detection which may be carried out in but a few seconds.
Heretofore leak testing using air as a testing medium required anywhere from 5 to 12 seconds for the air temperature in the test chamber being tested to stabilize before the leak test was effected. Without this stabilization, temperature changes of the air in the chamber being tested would give a false reading. Such temperature changes could result from the adiabatic cooling effect of compressed air expanding as it is released into the chamber, or, if the chamber walls were cooler or warmer than the air introduced into the chamber, heat transfers and consequent pressure changes would occur. With the requirement to increase production so that more parts could be tested in a given period of time, there has been the concurrent need to reduce the stabilization time.
Also, serious errors in leak detection based on the use of air as a test medium could arise in the production environment of a large manufacturing plant as a result of diurnal temperature variations which, in some instances, may be as much as 20.degree. F. in a 12-hour period. With this kind of variation, even a test taking less than one minute may permit a sufficiently significant temperature change in the test chamber to render wholly inaccurate a leak test based on air pressure change. Thus, there has been a need to neutralize these diurnal temperature effects which would cause changes in the temperature of the chamber being tested and distort the efficacy of the leak test.
Another error which can distort the accuracy of prior art leak testers may arise from the lack of uniformity of temperature of the parts being tested. For example, parts which have been waiting for some hours at room temperature to be leak tested might be cooler than parts coming directly from a parts washer, and yet if this temperature difference is not compensated for, there will be inaccuracies in the test results, viz., those parts which are warmer will tend initially to heat the air to the part temperature, and thereafter when the test is being performed the air will be cooling as the part cools and create a pressure drop simulating a leak when a leak may not, in fact, exist. Thus, the need to compensate for such errors has been an important problem to be solved in arriving at a satisfactory leak tester.
Because of the difficulties of leak testing structures, such as automobile radiators or fuel tanks, and occasionally engine blocks using air as a test medium, such products are often tested in large waterbath type testing machines relying on visual detection of air bubbles to signal leaks. As such a method is slow and relies on human eyesight, a faster and more reliable method has been needed for some time.
In our copending application filed Feb. 13, 1984, Ser. No. 579,701, we disclose a method and apparatus for leak testing a chamber wherein the gas pressure in the test chamber is compared with the gas pressure in a reference chamber, and the rate of change of the pressure difference is read as a function of leakage of the test chamber. The reference chamber is exposed to the same temperature effects as those in the test chamber, preferably by either forming the reference chamber as an adjunct of the part containing the test chamber, or by exposing the reference chamber directly to the gas in the test chamber. While this approach works satisfactorily in many instances, in others it is difficult or impossible to expose a reference chamber to the same temperature effects as the test chamber, and thus, in such instances this method is not feasible.